1: %\documentclass[12pt,preprint]{aastex}

2: \documentclass{emulateapj}

3: \usepackage{times}

4: 

5: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

6: % AUTHOR'S MACROS

7: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

8: \newcommand{\beq}{\begin{equation}}

9: \newcommand{\eeq}{\end{equation}}

10: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

11: % AUTHOR'S SYMBOLS

12: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

13: \newcommand{\gapprox}{$\stackrel {>}{_{\sim}}$}

14: \newcommand{\lapprox}{$\stackrel {<}{_{\sim}}$}

15: \def\alp{\mbox{$\alpha$}}

16: \def\farcd{\hbox{$.\mkern-4mu^\circ$}}

17: \def\farcm{\hbox{$.\mkern-4mu^\prime$}}

18: \def\farcs{\hbox{$.\!\!^{\prime\prime}$}}

19: \def\earth{\hbox{$\oplus$}}

20: \def\arcmin{\hbox{$^\prime$}}

21: \def\arcsec{\hbox{$^{\prime\prime}$}}

22: \def\solar{\mbox{$_{\normalsize\odot}$}}

23: \def\fs{\hbox{$.\!\!^{\rm s}$}}

24: \def\deg{\hbox{$^\circ$}}

25: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

26: \newcommand{\ttbs}{\char'134}

27: \newcommand{\AmS}{{\protect\the\textfont2

28:   A\kern-.1667em\lower.5ex\hbox{M}\kern-.125emS}}

29: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

30: \newcommand{\lsim}{\ \raise

31: -2.truept\hbox{\rlap{\hbox{$\sim$}}\raise5.truept\hbox{$<$}\ }}

32: \newcommand{\gsim}{\ \raise

33: -2.truept\hbox{\rlap{\hbox{$\sim$}}\raise5.truept\hbox{$>$}\ }}

34: \newcommand{\simsim}{\ \raise

35: -2.truept\hbox{\rlap{\hbox{$\sim$}}\raise5.truept\hbox{$\sim$}\ }}

36: \newcommand{\FeH}{\mbox{[Fe/H]}\,}

37: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

38: 

39: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

40: % add words to TeX's hyphenation exception list

41: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

42: \hyphenation{author another re-commend-ed Post-Script}

43: 

44: \slugcomment{Accepted for Publication in the Astrophysical Journal, 6

45: November 2008}

46: 

47: \righthead{A New Diagnostic Method for Stellar Stratification}

48: \lefthead{Gouliermis, de~Grijs, \& Xin}

49: 

50: \shorttitle{A New Diagnostic Method for Stellar Stratification}

51: \shortauthors{Gouliermis, de~Grijs, \& Xin}

52: 

53: \begin{document}

54:  

55: \title{A New Diagnostic Method for Assessment of Stellar Stratification

56: in Star Clusters}

57: 

58: \author{Dimitrios A. Gouliermis}

59: \affil{Max-Planck-Institut f\"{u}r Astronomie, K\"{o}nigstuhl 17,

60: D-69117 Heidelberg, Germany\\ dgoulier@mpia.de}

61: %\email{dgoulier@mpia.de}

62: 

63: \and

64: 

65: \and

66: \author{Richard de Grijs and Yu Xin}   

67: \affil{Department of Physics \& Astronomy, The University of Sheffield,

68: Hicks Building, Hounsfield Road, Sheffield S3 7RH, UK\\

69: R.deGrijs@sheffield.ac.uk; xinyu@bao.ac.cn}

70: 

71: %\author{Richard de Grijs}

72: %\affil{Department of Physics \& Astronomy, The University of Sheffield,

73: %Hicks Building, Hounsfield Road, Sheffield S3 7RH, UK}

74: %\affil{National Astronomical Observatories, Chinese Academy of Sciences,

75: %Beijing 100012, China\\ 

76: %R.deGrijs@sheffield.ac.uk}

77: %\email{R.deGrijs@sheffield.ac.uk}

78: %\and 

79: %\author{Yu Xin}

80: %\affil{Department of Physics \& Astronomy, The University of Sheffield,

81: %Hicks Building, Hounsfield Road, Sheffield S3 7RH, UK}

82: %\affil{National Astronomical Observatories, Chinese Academy of Sciences,

83: %Beijing 100012, China}

84: %\affil{Yunnan Astronomical Observatory, Chinese

85: %Academy of Sciences, Kunming, Yunnan 650011, China\\ xinyu@bao.ac.cn}

86: %\email{xinyu@bao.ac.cn}

87: 

88: \begin{abstract} 

89: We propose a new method for the characterization of stellar

90: stratification in stellar systems. The method uses the mean-square

91: radius (also called the {\sl Spitzer radius}) of the system as a

92: diagnostic tool. An estimate of the observable counterpart of this

93: radius for stars of different magnitude ranges is used as the {\em

94: effective radius} of each stellar species in a star cluster. We explore

95: the dependence of these radii on magnitude as a possible indication of

96: stellar stratification. This method is the first of its kind to use a

97: dynamically stable radius, and though seemingly trivial it has never

98: been applied before.  We test the proposed method using model star

99: clusters, which are constructed to be segregated on the basis of a Monte

100: Carlo technique, and on {\sl Hubble Space Telescope} observations of

101: mass-segregated star clusters in order to explore the limitations of the

102: method in relation to actual data. We conclude that the method performs

103: efficiently in the detection of stellar stratification and its results

104: do not depend on the data, provided that incompleteness has been

105: accurately measured and the contamination by the field population has

106: been thoroughly removed. Our diagnosis method is also independent of any

107: model or theoretical prediction, in contrast to the `classical' methods

108: used so far for the detection of mass segregation.

109: \end{abstract}

110: 

111: \keywords{Galaxies: star clusters -- Magellanic Clouds -- stellar

112: dynamics -- Methods: statistical}

113: 

114: \section{Introduction}

115: 

116: {\sl Mass segregation} or {\sl stellar stratification} is the

117: phenomenon, according to which the massive stars in a star cluster are

118: mostly concentrated toward its center. Due to dynamical evolution

119: massive stars move inwards to the center of the cluster through two-body

120: interactions. Due to the same process less massive stars may move

121: outwards, so that on average the system `evaporates' (e.g., Lightman \&

122: Shapiro 1978; Meylan \& Heggie 1997). This is more commonly known as

123: {\sl dynamical} mass segregation. However, mass segregation is also

124: being observed in dynamically young stellar systems, supposedly due to

125: the formation of massive stars closer to the center of the newly-born

126: system (e.g., Murray \& Lin 1996; Bonnell \& Davies 1998). This type of

127: mass segregation is usually referred to as {\sl primordial}.

128: 

129: Both types of stellar stratification affect the physical properties of

130: the cluster in a manner which allows us to develop diagnostic tools

131: for its detection and quantification. Such tools involve the

132: identification of changes in the surface density profiles and the mass

133: functions of the clusters, or of the core radii of stars in different

134: mass groups. Specifically, stellar stratification in a star cluster

135: can be detected from the surface density profiles of stars in

136: different brightness ranges. If $f$ is the stellar surface density

137: (for every brightness range) within a distance $R$ from the center of

138: the cluster, then $f(R)$ can be used as an indication of

139: stratification at the magnitude limit, where it changes significantly

140: (e.g., Scaria \& Bappu 1981; Sagar et al. 1988; Subramaniam et al.

141: 1993; Kontizas et al.  1998). The density profiles can be approximated

142: by $\log{f} \propto \gamma \log{R}$ (Elson et al. 1987) and the slope,

143: $\gamma$, can be used to quantify any significant change in

144: $f(R)$. Although this approximation was well designed, in practice

145: peculiarities of the density profiles toward the center of clusters

146: that are not completely spherical tend to produce large uncertainties

147: in $\gamma$, which hide any potential mass segregation (see, e.g.,

148: Gouliermis et al. 2004).

149: 

150: Another well-applied method is the construction of the luminosity or

151: mass function (LF, MF) of the cluster and the investigation of its

152: radial dependence (e.g., Pandey et al. 1992; King et al. 1995; Fischer

153: et al. 1998; Kontizas et al. 1998, de Grijs et al. 2002a,b). If the

154: cluster is indeed segregated, then the appearance of more massive

155: stars toward its center will result in shallower LFs and MFs toward

156: smaller radii (e.g., de Grijs et al. 2002a). However, this method has

157: several constraints imposed by the difficulties in the construction of

158: an accurate MF based on the use of theoretical mass-luminosity (ML)

159: relations. The use of different ML relations may result in significant

160: differences in the MF slopes obtained, independently of whether any

161: mass segregation is actually present (de Grijs et al. 2002b). In

162: addition, the investigation of any radial dependence of the LF and MF

163: slopes is not as straightforward as it seems, since its detection is

164: rather sensitive to the radial distance range being considered, i.e.,

165: where the slope is to be estimated (Gouliermis et al. 2004).

166: 

167: A rather interesting method is based on the use of the core radius of

168: the cluster as estimated for stars in specific magnitude (or mass)

169: ranges (e.g., Brandl et al. 1996; de Grijs et al. 2002b). Core radii

170: can be derived from the corresponding stellar number-density profiles

171: by fitting to the latter the models of Elson et al. (1987) and by

172: approximating them to the canonical King models for Galactic globular

173: clusters (King 1966). However, this method seems to be model-dependent

174: to a degree, which makes its application difficult. Another proposed

175: method is to measure the mean stellar mass within a given radius

176: (e.g., Bonnell \& Davies 1998; Hillenbrand \& Hartmann 1998), but this

177: diagnostic requires accurate knowledge of the stellar masses, which is

178: not always feasible (de Grijs et al. 2002b; Gouliermis et al. 2004).

179: 

180: For the methods discussed above, detecting stellar stratification

181: requires observations at high resolution, so that detailed radial

182: density distributions for the fainter magnitude ranges and complete

183: MFs in several annuli around the center of the cluster can be

184: constructed accurately. These methods can provide useful information

185: about the segregated masses as well as on the radius up to which the

186: cluster is shown to be segregated. However, it seems that all of these

187: methods are very sensitive to the quality of the observed data, as

188: well as to model assumptions. Moreover, all of the methods for the

189: detection of stellar stratification used thus far are closely

190: associated to the fitting of predefined functions to the data and

191: searching for parametric differences in these functions. One such

192: method, for example, involves the study of the stellar projected

193: density distribution of stars of different magnitudes; here, the

194: differences in the exponents of the slopes fitted are seen as proof of

195: mass segregation. A second method is based on the inherently imperfect

196: derivation of radius-dependent MFs, while yet another makes use of the

197: core radius of the system for different stellar groups, relying on

198: apparent structures, but it is extensively model-dependent. The

199: problem that all these studies suffer from is that {\sl it is very

200: difficult to compare their results among each other, thus leading to

201: extensive discussions about the nature and even the reality of stellar

202: stratification}.

203: 

204: In this paper we present a robust method for the detection of stellar

205: stratification. The simple application of this method can yield direct

206: information about the spatial extent (radial distribution) of stellar

207: groups in different magnitude (mass) ranges. It is based on the notion

208: of a dynamically stable radius of a star cluster, the so-called

209: {\sl `Spitzer radius,'} and leads to an observed {\sl `effective

210: radius.'} The diagnostic method for stellar stratification is

211: described in \S~2. The diagnostic process is applied to and tested on

212: a set of simulated spherical star clusters in \S~3, which are

213: constructed on the basis of Monte Carlo simulations. We thus validate

214: the use of the diagnostic and the unavoidable constraints. We then

215: apply the method to real data of two intermediate-age clusters in the

216: Magellanic Clouds observed with the {\sl Advanced Camera for Surveys}

217: (ACS) on-board the {\sl Hubble Space Telescope (HST)}, and we report on

218: our results in \S~4. A comparison of the results of our method with

219: those of the classical method of the MF dependence for the test

220: clusters is also presented in \S~4. Conclusions and a discussion on

221: the use of this diagnostic tool are summarized in \S~5.

222: 

223: %It should be kept in mind that we focus our presentation on the effect of

224: %{\sl luminosity segregation} or {\sl stratification}, which of course can

225: %be translated to mass segregation with the use of mass-luminosity

226: %relations, with all the related uncertainties. This presentation is also

227: %focused on the segregation of main-sequence stars in young star clusters. 

228: %Post-main-sequence populations are excluded from the stellar samples used

229: %in the following sections as their luminosities are not directly related

230: %to the underlying stellar masses.

231: 

232: \section{Diagnostic for Spatial Stellar Stratification}\label{s:method}

233: 

234: \subsection{The Spitzer Radius}\label{s:spitztheory}

235: 

236: The radius commonly known as the {\sl `Spitzer Radius'} of a star

237: cluster was introduced by L. Spitzer Jr. and his collaborators, who

238: used it as a distance indicator for different stellar subclasses in a

239: cluster (see, e.g., Spitzer 1958). Since gravity operates proportional

240: to the inverse square of the distance between a dynamical center and a

241: subject mass, if the latter represents a group of masses distributed

242: around the center (e.g., stars in a cluster), then the square distance

243: of this group can be replaced by the mean-square distance of its

244: members. Specifically, the gravitational force in a spherical stellar

245: system at a distance $r$ from its center, is (e.g., Spitzer 1958,

246: 1969)

247: \beq \frac{dV}{dr} = \frac{GM_{r}}{r^{2}} ,

248: \label{app-a1} \eeq 

249: where $V$ is the potential and $M_{r}$ the mass contained within a

250: radius $r$. King (1965) showed that considering only one single type

251: of stars with mass $m$ and if $S(x)dx$ is the number of stars in

252: a strip of width $dx$, then assuming a Plummer (1911) stellar

253: distribution, the potential is

254: \beq V(r)=-\frac{2Gm}{r}\int_{0}^{r}S(x)dx{\rm .} \label{app-a2}

255: \eeq 

256: Furthermore, Spitzer, investigating energy equipartition between two

257: different mass groups, notes that the term $r^{2}$ in Eq.

258: (\ref{app-a1}) may be expressed as the mean value of the square

259: distance of all stars in a single mass group, if they were the only

260: cluster members. Consequently, the square root of this value gives $r$

261: in Eq. (\ref{app-a2}). This radius, later called the `Spitzer radius,'

262: corresponds to the distance up to which the stars of this specific

263: mass group affect the gravitational field of the cluster as a whole.

264: 

265: Therefore, the Spitzer radius of a star cluster is defined as the

266: mean-square distance of the stars from its center, \beq r_{\rm

267: Sp}=\sqrt{\frac{\displaystyle\sum^{N}_{i=1} {r_{i}}^{2}}{N}} ,

268: \label{eq-01} \eeq where $r_{i}$ is the radial distance of the $i^{\rm

269: th}$ star and $N$ is the total number of stars. For a Maxwellian

270: distribution of a group of masses in a parabolic potential well, which

271: represents that of a globular cluster very well, the half-mass radius

272: of the system is almost equal to ($\sim$90\% of) the Spitzer radius

273: (Spitzer 1969). Consequently, considering the limitations in the

274: measurement of the half-mass radius of a cluster versus the direct

275: measurement of the Spitzer radius, the latter can be considered quite

276: important in relation to the dynamical status of the cluster, being

277: used as its characteristic radius.

278: 

279: As far as stellar stratification is concerned, in a star cluster

280: displaying mass segregation one should be able to observe the more

281: massive stars concentrated toward the center of the system by plotting

282: the position coordinates of the stars according to the their

283: magnitudes. The Spitzer radius, being a dynamically stable radius, can

284: be used for the parameterization of the spatial distribution of stars

285: in different brightness ranges. Based on this assumption, we propose

286: here a method according to which stellar stratification in a star

287: cluster can be quantified directly by estimating the corresponding

288: Spitzer radius for every group of magnitudes, using

289: Eq. (\ref{eq-01}). This simple approach provides information on both

290: {\sl where} in the cluster every stellar-luminosity group is confined,

291: and at {\sl which brightness} stellar stratification occurs.

292: 

293: \subsection{Spitzer Radius and Stellar Stratification}\label{s:stratheory}

294: 

295: We test the hypothesis outlined above on artificial populous spherical

296: clusters that we constructed based on the use of the Monte Carlo

297: technique. The proposed method is mainly designed for the identification

298: of stellar stratification in clusters in the Magellanic Clouds (MCs),

299: and therefore the simulated clusters are taken to have structural

300: parameters similar to those in the MCs. Results based on observations of

301: MCs clusters with {\sl HST} are particularly considered due to their

302: deepness and spatial resolution. Specifically, a sample of four star

303: clusters (NGC~1818, NGC~2004, NGC~2100, and NGC~330) observed by Keller

304: et al. (2000) with the {\sl Wide-Field Planetary Camera 2} (WFPC2) is

305: used as an initial guideline in our simulations, by considering their

306: structural parameters for the artificial star clusters we construct

307: here. As a consequence, our simulated clusters have tidal radii of

308: $r_{\rm t}\simeq$ 2\farcm0, while King models with $C=\log{r_{\rm

309: t}/r_{\rm c}} \sim 1.0$ are assumed to represent their density profiles.

310: The corresponding core radii, $r_{\rm c}$, of the artificial clusters

311: have values within the limits given by Mackey \& Gilmore (2003a,b), who

312: compiled two-color {\sl HST} observations for a sample of 53 and 10 rich

313: star clusters in the Large and Small Magellanic Cloud (LMC, SMC),

314: respectively.

315: 

316: The LFs of the simulated clusters are selected as variations of a

317: global LF, which was constructed based on the observed LFs of the

318: clusters in the sample of Gouliermis et al. (2004). This global LF is

319: found to be in very good agreement with the LFs of star clusters in

320: the MCs as determined using {\sl HST} imaging by other authors (e.g., de

321: Grijs et al. 2002c: NGC~1805, NGC~1818, NGC~1831, NGC~1868, NGC~2209,

322: and Hodge~14; Sirianni et al. 2002: NGC~330; Testa et al. 1999,

323: Brocato et al. 2003: NGC~1866; Santiago et al. 2001: NGC~1805,

324: NGC~1818, NGC~1831, NGC~1868, and NGC~2209). An example of three

325: artificial clusters with different LFs is given in

326: Fig.~\ref{f:mcsample}. The LF in each case is also given at the bottom

327: panel of the figure. All three clusters have the same tidal radius and

328: the same degree of stratification, but the differences in their LFs

329: are fed through to their stellar numbers and consequently to their

330: appearance.

331: 

332: Stellar stratification was considered for our artificial star

333: clusters.  Hence, several degrees of segregation (including no

334: segregation) were adopted. This was done using a projected limiting

335: radius within which each stellar group should be confined. Based on

336: our hypothesis, stellar stratification in these clusters should be

337: detected through the dependence of the estimated Spitzer radius of

338: stars in specific magnitude ranges versus the corresponding mean

339: magnitude. Indeed, our tests show that the initially assumed stellar

340: stratification (of any degree) is reconstructed by the plot of Spitzer

341: radii per magnitude range versus magnitude for all artificial star

342: clusters considered. In Fig.~\ref{f:rspitzsmpl} we show a sample of

343: three degrees of stratification for the same simulated cluster

344: (similar to the middle panel in Fig.~\ref{f:mcsample}). Spitzer radii

345: in this figure are normalized for reasons of comparability. Three

346: different symbols were used for the plot, as if the cluster was

347: observed projected onto the $xy$, $xz$, or $yz$ plane, respectively.

348: 

349: \subsection{Spitzer Radius of Observed Star Clusters}

350: 

351: Identical to the concept discussed in the previous sections,

352: observations of the radial distribution of stars in a real star cluster

353: give us a relation of the kind of Eq. (\ref{eq-01}), \beq r_{\rm

354: obs}=\sqrt{\frac{\displaystyle\sum^{N_{\rm obs}}_{i=1}

355: {r_{i}}^{2}}{N_{\rm obs}}} . \label{eq-02} \eeq However, this

356: distribution is significantly affected by two important observational

357: constraints: (i) incompleteness of the stellar sample and (ii)

358: contamination by field stars. To obtain an accurate measurement of the

359: observed Spitzer radii per magnitude range in a cluster, both these

360: constraints should be considered.

361: 

362: \subsubsection{Incompleteness of the observations}

363: 

364: The observed stellar samples in star clusters are incomplete due to the

365: observations. Although this can lead to a parameterization significantly

366: differing from reality, the bias introduced to the data analysis is, in

367: general, well understood. Consequently, solving the problem of

368: incompleteness is a more or less straightforward procedure, which takes

369: place through extensive artificial-star tests (e.g. de~Grijs et al.

370: 2002a; Gouliermis et al. 2004). A set of artificial stars is generated

371: within each of the observed frames.  Then, an identical reduction

372: procedure is performed on the artificially enriched frames, in order to

373: estimate the number of the `recovered' artificial stars. The

374: completeness factor $C$ is the ratio between the number of stars

375: recovered to the number of stars originally simulated.  This factor

376: depends on the brightness of the stars and their positions in the

377: cluster. Therefore, an effective completeness estimation for a stellar

378: cluster requires that incompleteness is calculated for stars in

379: different magnitude bins and for different distances from the center of

380: the cluster.

381: 

382: In order to correct the observed Spitzer radius for incompleteness, for

383: every counted star its completeness factor is calculated according to

384: its magnitude and radial distance. If the radial distance of the $i^{\rm

385: th}$ star is $r_{i}$ and its corresponding completeness factor is

386: $C_{i}$ then we assign $1/C_{i}$ stars to this distance and the total

387: number of counted stars is affected accordingly. Thus, the

388: completeness-corrected observed Spitzer radius should be \beq r_{\rm

389: obs,cc}=\sqrt{\frac{\displaystyle\sum_{i=1}^{N_{\rm

390: obs}}({r_{i}}^{2}/C_{i})} {\displaystyle\sum_{i=1}^{N_{\rm

391: obs}}(1/C_{i})}} . \label{eq-03} \eeq

392: 

393: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

394: \begin{figure*}[t!]

395: \epsscale{1.15} 

396: \plotone{figure1.eps}

397: \caption{Sample of three different simulated star clusters using Monte

398: Carlo simulations. The maps of the star clusters, which differ in

399: their luminosity functions, are shown in the top panel. The

400: corresponding LFs are shown in the bottom panel. All clusters were

401: assumed to follow a King density profile and to have the same tidal

402: radius of 0\farcm2. They are also similarly segregated. The LFs shown

403: are in agreement with the results of various authors on star clusters

404: in the MCs.\label{f:mcsample}}

405: \end{figure*}

406: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

407: 

408: \subsubsection{Contribution from the field populations.}

409: 

410: The contamination of the observations by the field populations is

411: probably the most difficult problem to be dealt with. Theoretically,

412: if even one massive field star was counted as a cluster member and if

413: it is far away from the center of the cluster, then the measured

414: radius would be overestimated. In the same context, with comprehensive

415: observations of the field we know the expected number of contaminating

416: stars per magnitude range, but we still cannot know their position,

417: which is the most important constraint as far as the radius estimation

418: is concerned.

419: 

420: One may confront this problem by first defining the largest radial

421: distance among the stars of the entire sample in a specific magnitude

422: bin and then normalizing the known surface density of field stars in

423: this magnitude range to the surface confined by this radius. Hence,

424: one could obtain a good estimation of the number of field stars

425: expected to contaminate this area in the specific magnitude range, and

426: correct for field contribution by subtracting the appropriate fraction

427: of the total number of field stars. If, for example, there are $N_{\rm

428: F}$ field stars expected in a total number, $N$, of stars in a

429: specific magnitude range (within a maximum radial distance, $r_{\rm

430: max}$), then we can subtract their fraction $F = N_{\rm F}/N < 1$ from

431: every counted star with distance $r_{\rm i} \leq r_{\rm max}$.  Then,

432: the field-subtracted observed Spitzer radius, ignoring for the moment

433: any incompleteness corrections, derived from Eq.  (\ref{eq-02}) for

434: stars in every magnitude range should be

435: \beq 

436:  r_{\rm obs,fs} = \sqrt{\frac{\displaystyle\sum_{i=1}^{N_{\rm obs}} \left[

437: (1 - {F}){r_{i}}^{2}\right]}{N_{\rm obs} - N_{\rm F}}} . \label{eq-04}

438: \eeq

439: 

440: However, this simplified approach assumes that the field stars are

441: more or less homogeneously distributed within the boundaries of the

442: cluster, which does not correspond to reality. Specifically,

443: Eq.~(\ref{eq-04}) weighs the contribution of each star to the sum of

444: ${r_{i}}^{2}$ by the probability, $(1-F)$, that it is a cluster

445: member.  Still, cluster members are by their very nature more

446: centrally concentrated, and therefore stars at smaller radii are less

447: likely to be field stars, while those at large radii have a higher

448: probability that they are, in fact, field stars. As a consequence, in

449: Eq.~(\ref{eq-04}) true cluster members make a smaller contribution to

450: the Spitzer radius than they should, while the contribution of the

451: field stars is overestimated. We verified this problem using

452: additional Monte Carlo simulations for the construction of

453: hypothetical background-field populations. For these simulations, the

454: use of a specific field LF was the only constraint considered.

455: 

456: %\beq r_{\rm obs,fs} = \sqrt{\frac{\displaystyle\sum_{i=1}^{N_{\rm

457: %obs}}{r_{i}}^{2} - \sum_{i=1}^{N_{\rm obs}}{F}{r_{i}}^{2}}{N_{\rm obs} -

458: %N_{\rm F}}} \equiv \eeq

459: 

460: We constructed homogeneously distributed stellar populations with LFs

461: typical for the background fields of both Magellanic Clouds, based on

462: {\sl HST} observations (e.g., Elson et al. 1997; Holtzman et al. 1999;

463: Dolphin et al. 2001; Smecker-Hane et al. 2002). Subsequently, we

464: inserted our artificial star clusters (\S~\ref{s:stratheory}) into the

465: artificial background field. We constructed 32 different cases of four

466: different clusters, each having four different degrees of stratification

467: located in two different types of field (with different LFs). The

468: application of our method to these artificial observations showed that

469: {\sl it is indeed very difficult to disentangle the contribution of the

470: field} in the measured Spitzer radii on the basis of Eq.~(\ref{eq-04}).

471: Consequently, and since it is not possible to quantify the effect of the

472: field contribution to observed stellar samples in clusters as a function

473: of distance from the cluster center, a more robust approach for the

474: direct field decontamination of the observed stellar samples {\sl before

475: measuring the Spitzer radii} should be considered. Such a

476: well-established method for the removal of any field contamination from

477: the stellar samples of star clusters is based on the application of a

478: Monte Carlo technique, which makes use of comprehensive observations of

479: the local background field of the galaxy (e.g., Bonatto \& Bica 2007).

480: We apply such a sophisticated technique to actual {\sl HST}/ACS

481: observations of two SMC clusters, for the estimation of their Spitzer

482: radii as a function of stellar magnitude, in \S~3.

483: 

484: %\beq

485: %r_{\rm obs,cc,fs} =

486: %\sqrt{\frac{\displaystyle\sum_{i=1}^{N_{\rm obs}}({r_{i}}^{2}/C_{i})  -

487: %\sum_{i=1}^{N_{\rm obs}}{F}{r_{i}}^{2}}{\displaystyle\sum_{i=1}^{N_{\rm

488: %obs}}(1/C_{i}) -N_{\rm F}}}

489: %\label{eq-03}

490: %\eeq

491: 

492: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

493: \begin{figure}[t!]

494: \epsscale{1.15} 

495: \plotone{figure2.eps}

496: \caption{Sample of three different degrees of segregation: (a) No

497: stellar stratification, (b) smooth stratification, and (c) for more

498: strongly segregated bright stars. The artificial cluster shown in the

499: middle panel of Fig. \ref{f:mcsample} is used for this demonstration.

500: The brightest (segregated) stars appear to be confined within smaller

501: Spitzer radii than the faint (non-segregated) stars.  Different

502: symbols for each magnitude range represent $r_{\rm Sp}$ as it would be

503: estimated if the cluster were observed projected onto the $xy$, $xz$,

504: or $yz$ plane, respectively.\label{f:rspitzsmpl}}

505: \end{figure}

506: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

507: 

508: 

509: \subsection{The Effective Radius}

510: 

511: The method we propose for the diagnosis of stellar stratification in

512: star clusters makes use of the observable counterpart of Spitzer

513: radius, applied to the observed completeness-corrected stellar

514: samples, and after the contamination of these samples by the field

515: populations has been removed. This radius, which we will refer to as

516: the {\sl effective radius}, $r_{\rm eff}$, is estimated for stars in

517: specific narrow magnitude ranges; its dependence on magnitude is shown

518: in the following sections. Eq. (\ref{eq-03}) can be reduced to the

519: following expression and should, with all corrections properly

520: applied, be equal to $r_{\rm Sp}$. Thus, \beq r_{\rm

521: eff}=\sqrt{\frac{\displaystyle\sum_{i=1}^{N_{\rm

522: obs,cc,ff}}{\frac{{r_{i}}^{2}}{C_{i}}} }{N_{\rm obs,cc,ff}}} \equiv

523: r_{\rm Sp} \label{eq-07} , \eeq where $r_{i}$ is the radial distance

524: of the $i^{\rm th}$ stellar member of the cluster (after field

525: subtraction), in a specific magnitude range. $C_{\rm i}$ is the

526: corresponding completeness factor at its distance for the same

527: magnitude range and $N_{\rm obs,cc,ff}$ is the total number of

528: observed stars corrected for incompleteness and reduced by the total

529: number of field stars.

530: 

531: This radius is a statistical quantity and its estimation is rather

532: sensitive to the number of observed stars. Therefore, the effective

533: radius can be easily over- or underestimated if this estimation is

534: based on few stars only, which might be the case for the brightest

535: magnitude ranges. In order to deal with this issue, one may apply the

536: binning in variable magnitude ranges and use wider magnitude bins

537: toward the brighter limit, thus increasing the number of the sample

538: stars. However, with this solution one may not be able to observe the

539: luminosity segregation of the stars in detail. It would be more

540: appropriate to weigh each bin by the number of sample stars as an

541: indicator of the statistical significance of the estimation of the

542: corresponding effective radius.

543: 

544: The uncertainty in the estimation of $r_{\rm eff}$, as derived from

545: Eq. (\ref{eq-07}), is mostly dependent on errors in the counting

546: process, thus reflecting Poisson statistics. Another concern to be

547: taken into account is the effect of projection on the estimated values

548: for the effective radius. For non-symmetrical loose stellar systems,

549: it is found that projection does not significantly affect the true

550: value of the Spitzer radius (e.g., Gouliermis et al. 2000). Here,

551: since we deal with spherically symmetric clusters, we consider the

552: uncertainties introduced by projection for the simple case of a

553: homogeneous density distribution.

554: 

555: We assume a spherical distribution of $N$ particles ($N \gg 1$),

556: uniformally distributed (with constant density, $\varrho$) within a

557: radius, $R$. If one observes this sphere, the distribution is not

558: uniform anymore because of projection effects.  Specifically, we

559: investigate the size of the radius which includes half of the

560: particles, projected along the line of sight, in relation to the

561: radius of a sphere which -- in reality -- includes half of them (in 3D

562: space). Because of symmetry we concentrate on one hemisphere, defined

563: by the line of sight (direction of the Y axis in

564: Fig.~\ref{f:projfig}).

565: 

566: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

567: \begin{figure}[t!]

568: \epsscale{1.15} 

569: \plotone{figure3.eps}

570: \caption{Projection of a hemispherical distribution. The $Y$ axis coincides 

571: with the line of sight.\label{f:projfig}}

572: \end{figure}

573: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

574: 

575: We define an elementary cylinder of thickness $dx =

576: R~\cos{\theta}~d\theta$, of base radius $x = R~\sin{\theta}$, and

577: of height $y = R~\cos{\theta}$ (Fig.~\ref{f:projfig}). The angle

578: $\theta$ is measured in relation to the radius vertical to the surface

579: of projection. The cylinder contains $dN = \varrho~dV$

580: particles, where $V$ is the volume. Consequently,

581: \beq

582: dN\left ( \theta \right ) = 2\pi~R^{3}~\varrho~\sin{\theta}~ 

583: \cos^{2}\mkern-4mu{\theta}~d\theta .

584: \eeq

585: 

586: Integrating over $0{\deg} < \theta < 90{\deg}$ we expect that the

587: number of particles included is $N = 2\pi~R^{3}~\varrho/3$. Half

588: ($N/2$) are projected up to a limit cylinder base radius, $x_{\rm h}$,

589: corresponding to $\theta_{\rm h}$. This angle can be deduced from the

590: relation \begin{displaymath} N/2 = \int_{0}^{\theta_{\rm h}} {\rm

591: d}N\left ( \theta \right) = -2 \pi R^{3} \varrho \int_{0}^{\theta_{\rm

592: h}} \cos^{2}\mkern-4mu{\theta}~d \left ( \cos{\theta} \right )

593: ~\Rightarrow \end{displaymath} \beq \Rightarrow

594: ~\cos^{3}\mkern-4mu{\theta_{\rm h}} = \frac{1}{2} ~\Rightarrow~

595: \theta_{\rm h} \approx 37\farcd467 . \eeq This angle gives $x_{\rm h}

596: = R~\sin{\theta_{\rm h}} \approx 0.608R$. In reality, half of the

597: particles are included in a volume contained within a sphere of radius

598: $R_{\rm h} = 2^{-1/3}R \simeq 0.794R$ (because of constant

599: density). Thus, we have $R_{\rm h} = 1.3x_{\rm h}$.  As a consequence,

600: the proportion of the half-number radius (and consequently of the

601: Spitzer radius, which is almost equal to the half-number radius for a

602: relaxed stellar system) to its projected value for this distribution

603: of particles is well-defined.

604: 

605: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

606: \begin{figure}[t!]

607: \epsscale{1.15} 

608: \plotone{figure4.eps}

609: \caption{Completeness functions for Kron~1 (solid line) and Lindsay~91

610: (dashed line), respectively. The function has been integrated over

611: position and color as a function of m$_{\rm F555W}$.\label{f:compl}}

612: \end{figure}

613: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

614: 

615: In any case, in the present study this factor is irrelevant for

616: further analysis, since the application of a constant conversion

617: factor to the estimated effective radii is not necessary. Moreover,

618: this application assumes that any cluster follows a homogeneous

619: density distribution, which is of course not always the case. Under

620: these circumstances, we can treat the estimated effective radii used

621: in the application of our method as the {\sl projected} effective

622: radii.

623: 

624: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

625: \begin{figure*}[t!]

626: \epsscale{1.15} 

627: \plotone{figure5.eps}

628: \caption{m$_{\rm F555W}$$-$m$_{\rm F814W}$ vs. m$_{\rm F555W}$ CMDs of 

629: the fields of the SMC clusters Kron~1 and Lindsay~91, observed with

630: {\sl HST} ACS/WFC.\label{f:acscmds}}

631: \end{figure*}

632: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

633: 

634: \section{The Star Cluster Sample}\label{s:data}

635: 

636: The proposed method for the detection of stellar stratification is

637: designed primarily for the study of MC clusters, and therefore in the

638: following sections we apply this method to such clusters to test its

639: performance and to establish a consistent methodology. Moreover, in

640: order to use the most complete available data we select clusters

641: observed with the {\sl HST}. The advantages {\sl HST} introduced to

642: crowded-field photometry in the MCs have been documented extensively

643: by several authors since its early observations and the quality of the

644: data obtained with both the Wide-Field Planetary Camera 2 (WFPC2) and

645: the Advanced Camera for Surveys (ACS) has proven to be more than

646: adequate for detailed studies of young and old MC star clusters (e.g.,

647: Fischer et al. 1998; Johnson et al. 1999; Keller et al. 2000; Brocato

648: et al. 2001; de~Grijs et al. 2002a,b,c; Stanghellini et al. 2003;

649: Gouliermis et al. 2004; Mackey et al. 2006; Rochau et al. 2007; Xin et

650: al. 2008).

651: 

652: For the application of our new method for diagnosis of stellar

653: stratification to real clusters we selected a set of data obtained with

654: the ACS of the intermediate-age clusters Kron~1 (Kron 1956) and

655: Lindsay~91 (Lindsay 1958) in the SMC. Below, we discuss the reduction of

656: the observations of these clusters (see also Mackey \& Gilmore 2004),

657: their photometry, and the decontamination of the cluster populations

658: from the contribution of the general background field of the galaxy

659: using an advanced Monte Carlo technique.

660: 

661: \subsection{Data Reduction}

662: 

663: The SMC clusters Kron~1 and Lindsay~91 were observed with the Wide

664: Field Channel (WFC) of ACS on-board the {\sl HST} (program GO-9891, PI

665: G. Gilmore). The ACS/WFC consists of two 2048$\times$4096-pixel CCDs,

666: separated by a gap of $\sim$50 pixels.  It covers a field of view of

667: 202$\times$202 arcsec$^2$, with a scale of $\sim$0.05 arcsec

668: pixel$^{-1}$.  The frames were taken in each of the F555W and F814W

669: filters.  Exposure times were 300 and 200 seconds, respectively. More

670: details on the observations and instruments can be found in Mackey \&

671: Gilmore (2004).

672: 

673: The {\sc fits} files of the two clusters were retrieved from the Space

674: Telescope Science Institute (STScI) Data Archive. The original data

675: were reduced with the STScI pipeline, i.e., they had been bias and

676: dark-current subtracted and divided by flat-field images.  Photometry

677: was performed with the ACS module of the package {\sc

678: dolphot}\footnote{The ACS module of {\sc dolphot} is an adaptation of

679: the photometry package {\sc HSTphot} (Dolphin 2000). It can be

680: downloaded from {\tt http://purcell.as.arizona.edu/dolphot/}.}

681: (Version 1.0). Photometry processes and corresponding parameters fully

682: follow the procedures and recommendation in the {\sc dolphot} manual.

683: Photometric calibrations and transformations were done using the

684: relations in Sirianni et al. (2005).

685: 

686: We adopted three parameters in {\sc dolphot} to filter the photometric

687: results, i.e., we selected only objects with

688: $-$0.3~$\leq$~sharpness~$\leq$~0.3, crowding~$\leq$~0.5~mag, and

689: $\chi^2$~$\leq$~0.25 in both frames. Meanwhile, we kept objects

690: classified as good star (type 1) and star errors of types 1 to 7 by {\sc

691: dolphot}, which are referred to as `usable' in the {\sc dolphot} manual.

692: To calculate the completeness of the photometry, {\sc dolphot} was run

693: again in the artificial-star mode. For each cluster, we generated

694: $\sim10^6$ fake stars with limits of 16.0~-~28.0~mag in brightness and

695: $-$0.50~-~2.00~mag in color. The effect of unrealistically large crowding

696: that would lead to over-estimation of the incompleteness was taken into

697: account by the {\sc dolphot} utility {\sl acsfakelist}, which creates

698: artificial star lists based on the original photometric catalog. The

699: fake stars were binned in four dimensions, i.e, $x$ and $y$ positions,

700: magnitude, and color. Fig.~\ref{f:compl} presents the completeness

701: function of Kron~1 (solid line) and Lindsay~91 (dashed line),

702: respectively. The function is integrated over position and color and is

703: a function of m$_{\rm F555W}$.  More details on the data reduction are

704: given in Xin et al. (2008).

705: 

706: \subsection{Field Decontamination}\label{ss:fd}

707: 

708: The observed color-magnitude diagrams (CMDs) of the ACS fields centered

709: on Kron~1 and Lindsay~91 are shown in Fig.~\ref{f:acscmds}. Both fields

710: suffer from significant contamination by field stars, as can also be

711: seen in the corresponding maps shown in Fig.~\ref{f:acsmaps}. Without

712: any information about cluster-membership probabilities for the observed

713: stars, obtained, e.g., from radial velocities and/or proper motions, the

714: quantitative decontamination of the clusters from the field stars on a

715: statistical basis becomes a fundamental method to obtain the CMD of the

716: true stellar members of each cluster (see e.g. the pioneering work by

717: Flower et al. 1980 on the LMC cluster NGC~1868).

718: 

719: In this work, the algorithm described in Bonatto \& Bica (2007) is

720: used for field-star decontamination of the observed CMDs of both

721: clusters.  As a first step, we use the stellar number-density profiles

722: of the clusters to identify the radial distance from the center of

723: each cluster, R$_{\rm lim}$, where the stellar density becomes flat,

724: indicating that the background-field density has been reached. We call

725: this region the `offset region' of the cluster. All stars in the

726: offset regions (with $r>R_{\rm lim}$) are treated as field stars,

727: while the rest (with $r \leq R_{\rm lim}$) are considered as the most

728: probable cluster-member stars.  Assuming a homogeneous field-star

729: distribution, the number density of field stars is applied to the

730: whole cluster region to remove the field contamination. We perform

731: this calculation in two dimensions in the CMD, i.e., in m$_{\rm

732: F555W}$ and in (m$_{\rm F555W}$$-$m$_{\rm F814W}$), considering also

733: the observational uncertainties in the photometry, $\delta_{\rm

734: F555W}$ and $\delta_{\rm F814W}$.

735: 

736: In short, the decontamination process is done as follows: (i) we

737: divide the CMDs of the cluster and offset-region stars, respectively,

738: in 2D cells with the same axes along the m$_{\rm F555W}$ and (m$_{\rm

739: F555W}$$-$m$_{\rm F814W}$) directions, (ii) we then calculate the

740: expected number density of field stars in each cell in the CMD of the

741: offset region, and (iii) we randomly subtract the expected number of

742: field stars from each cell from the CMD of the cluster region. In the

743: following description we use the symbols $\chi$=m$_{\rm F555W}$ and

744: $\xi$= (m$_{\rm F555W}$-m$_{\rm F814W}$) to simplify the

745: notation. Consider a CMD cell with sides and coordinates

746: $\chi_c\pm\triangle\chi/2$ and $\xi_c\pm\triangle\xi/2$, respectively,

747: where ($\chi_c$, $\xi_c$) are the cell's central coordinates. We

748: assume a Gaussian probability distribution to calculate the

749: probability of a star with CMD coordinates

750: ($\bar{\chi}\pm\delta_\chi$, $\bar{\xi}\pm\delta_\xi$), with

751: $\delta_\xi$=($\delta^2_{\rm F555W}$+$\delta^2_{\rm F814W}$)$^{1/2}$,

752: to be found in the cell, and therefore the computations take into

753: account both magnitude and color uncertainties, for example,

754: 

755: \beq

756: P(\chi,\bar{\chi})=\frac{1}{\sqrt{2\pi}\delta_\chi}e^{(-1/2)\left[(\chi-\bar{\chi})/\delta_\chi\right]^2} .

757: \eeq

758: 

759: The expected field-star number density ($\rho^{\rm cell}_{\rm fs}$) in

760: a cell is given by summing up the individual probabilities ($P^{\rm

761: cell}_{\rm fs}$) of all offset-field stars (N$_{\rm fs}$) in the cell,

762: divided by the offset area (A$_{\rm fs}$), i.e., $\rho^{\rm cell}_{\rm

763: fs}$=$P^{\rm cell}_{\rm fs}$/A$_{\rm fs}$, where

764: 

765: \beq

766: P_{\rm fs}^{\rm cell}=\sum_{i=1}^{N_{\rm fs}}\int\int P_i(\chi, \chi_i; \xi, \xi_i) d\chi d\xi .

767: \eeq

768: 

769: $P_i(\chi, \chi_i; \xi, \xi_i)$ is the probability of the

770: \textit{i}$^{\rm th}$ field star, with CMD coordinates ($\bar{\chi}$,

771: $\bar{\xi}$) and uncertainties ($\delta_{\chi i}$, $\delta_{\xi i}$), to

772: have the magnitude and color ($\chi$, $\xi$).  The integral is carried

773: out in the two dimensions, $\chi_c$: $\triangle \chi/2 \leq \chi \leq

774: \chi_c + \triangle \chi/2$, and $\xi_c$: $\triangle \xi/2 \leq \xi \leq

775: \xi_c + \triangle \xi/2$, respectively.  In the same manner we calculate

776: the number density of the observed stars in the cell in the

777: cluster-region CMD, $\rho^{\rm cell}_{\rm obs}=P^{\rm cell}_{\rm

778: obs}$/A$_{\rm cl}$, where A$_{\rm cl}$ is the projected area of the

779: cluster region ($r \leq R_{\rm lim}$). Therefore, the expected number of

780: field stars in the cell in the cluster-region CMD is $n^{\rm cell}_{\rm

781: fs} = (\rho^{\rm cell}_{\rm fs}/\rho^{\rm cell}_{\rm obs}) \times n^{\rm

782: cell}_{\rm obs}$, where $n^{\rm cell}_{\rm obs}$ is the number of

783: observed stars (at $r\leq R_{\rm lim}$) located in the cell. The number

784: of probable cluster member stars in the cell will be $n^{\rm cell}_{\rm

785: cl} = n^{\rm cell}_{\rm obs}-n^{\rm cell}_{\rm fs}$, and the total

786: number of cluster member stars is the sum of all $n^{\rm cell}_{\rm

787: cl}$, i.e., $N_{\rm cl}=\sum_{\rm cell}n^{\rm cell}_{\rm cl}$. 

788: 

789: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

790: \begin{figure*}[t!]

791: \epsscale{1.15} 

792: \plotone{figure6.eps}

793: \caption{The observed ACS/WFC fields of view of Kron~1 (left) and

794: Lindsay~91 (right).  The overlayed circles have radii, $R_{\rm lim}$,

795: corresponding to the boundary between the clusters and the general

796: field. They were defined based on the clusters' surface density

797: profiles, as the distance from the cluster center where the density

798: profiles become flat (see \S~\ref{ss:fd}).\label{f:acsmaps}}

799: \end{figure*}

800: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

801: 

802: To minimize any artificial effect intrinsic to the method, we apply the

803: algorithm many times using different cell sizes ($\triangle\chi$,

804: $\triangle\xi$) and derive different decontamination results, from which

805: we calculate the probability of a star to be identified as a `true'

806: cluster member. Finally the field-star-decontaminated CMD of the

807: cluster contains the $N_{\rm cl}$ stars with the highest probability of

808: being cluster members.  Fig.~\ref{f:fsdcmds} shows the corresponding

809: CMDs of the cluster region (left panel), the offset region (central

810: panel), and the final field-star-decontaminated CMD (right panel) for

811: both clusters.

812: 

813: \section{Application of the diagnostic method}\label{s:application}

814: 

815: The proposed method for the diagnosis of stellar stratification is

816: based on the assumption that the {\sl effective radius} of stars in a

817: specific magnitude (or mass) range will be a unique function of this

818: range if the system is segregated. In this section we apply our new

819: method for the detection and quantification of stellar stratification

820: to the intermediate-age clusters Kron~1 and Lindsay~91. Before the

821: application of our diagnostic method and to check if the clusters are

822: segregated, we investigate the stellar stratification using one of the

823: most accurate `classical' methods, the radial dependence of the

824: cluster LFs and MFs.

825: 

826: In the previous section we showed that both clusters are evolved, by

827: virtue of the presence of clear red-giant branches (RGBs) and red

828: clumps in their CMDs. However, in the corresponding `clean' CMDs,

829: obtained after decontamination of the cluster stellar samples by the

830: field populations, shown in Fig.\ref{f:fsdcmds} (right panels), it can

831: be seen that there are stars at the upper main sequence (MS) and just

832: above the MS turn-off, which are not sufficiently removed from the

833: observed CMDs. These stars are most probably members of the younger

834: SMC field, to which both clusters belong, and they will not be

835: considered in the following analysis. As a consequence, for the

836: investigation presented here, we use only the lower-MS population

837: below the turn-off of each cluster and that of the RGB.

838: 

839: \subsection{Stellar Stratification in the selected clusters}

840: 

841: In this section we apply the most effective {\sl classical} diagnostic

842: method developed in the past for the investigation of stellar

843: stratification to our ACS photometry of Kron~1 and Lindsay~91, to

844: check if the clusters are indeed segregated. This well-applied method

845: requires the construction of the cluster LF and/or the MF and the

846: investigation of its radial dependence. If the cluster is indeed

847: segregated then the appearance of more massive stars toward its center

848: will result in a shallower MF at smaller radii. De Grijs et al.

849: (2002b) present the difficulties in the construction of an accurate MF

850: based on the use of different ML relations and stress the differences

851: in the MF slopes obtained. Moreover, significant numerical biases in

852: the determination of the slope of the MF using linear regression are

853: found if the construction of the MF was made from uniformly binned

854: data (Ma{\'{\i}}z Apell{\'a}niz \& {\'U}beda 2005), implying the

855: presence of systematic errors in the slopes of MFs calculated in this

856: way.

857: 

858: Gouliermis et al. (2004) argue that the investigation of any radial

859: dependence of the MF slope is not as straightforward as it seems,

860: since any such dependence is rather sensitive to the selection of the

861: radial distances being considered. Although stellar stratification can

862: be identified from the cluster LF, its quantification is quite

863: difficult because the approximation of the LF with a single power law,

864: which is the usual approach, can only give very rough results. In any

865: case, we apply this diagnostic to our data, in order to identify

866: stellar stratification in our clusters and to be able to compare the

867: results of a well-established classical method with those of our

868: method.

869: 

870: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

871: \begin{figure*}[t!]

872: \epsscale{1.15} 

873: \plotone{figure7.eps}

874: \caption{CMDs of the cluster regions ($r \leq R_{\rm lim}$; left panel),

875: the offset regions ($r > R_{\rm lim}$; central panel), and the final CMDs

876: of the clusters after field-star decontamination (right panel) for both

877: Kron~1 (top) and Lindsay~91 (bottom).\label{f:fsdcmds}}

878: \end{figure*}

879: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

880: 

881: 

882: \subsubsection{Radial Dependence of LFs and MFs of the Clusters}

883: 

884: In this section we examine the dependence of the shape and slope of

885: the stellar LF on position within the clusters. We basically use the

886: cluster LFs as smoothing functions of the full 2D ($m_{\rm F555W},

887: m_{\rm F555W} - m_{\rm F814W}$) CMDs shown in the right-hand panels of

888: Fig. \ref{f:fsdcmds}. In Fig. \ref{f:lfs} we show all annular cluster

889: LFs out to $r = 100.0''$, corrected for the effects of incompleteness

890: and background contamination at each respective annular radius. We

891: carefully considered the optimal radial ranges to be used for our LF

892: analysis, which was in essence driven by the need to have

893: statistically significant {\it and} similar (cf. Ma{\'{\i}}z

894: Apell{\'a}niz \& {\'U}beda 2005) numbers of stars in each of the

895: subsamples used for our comparison of the different annular LFs. For

896: reasons of clarity, we did not normalize the LFs to the same sampling

897: area since this would result in smaller separations among the various

898: curves, hence hampering our assessment of any radial dependence of the

899: LF {\it shape} (at this point we are not interested in the absolute

900: numbers of stars per unit area).

901: 

902: We note that for a proper comparison of the annular LFs in Fig.

903: \ref{f:lfs} one should only consider the magnitude range in between the

904: vertical dotted lines. For brighter stars, stochasticity in the stellar

905: LFs starts to dominate, while the LFs are also significantly affected by

906: the presence of red-clump stars, which are clearly visible in both CMDs.

907: For fainter stars our subsamples are significantly statistically

908: incomplete, at $<$50\%. On the basis of a casual comparison of the

909: annular LFs in Fig. \ref{f:lfs}, we conclude that there is little

910: evidence for luminosity segregation in Kron 1, although there is a hint

911: of a flattening of the LFs towards smaller radii, particularly for

912: $m_{\rm F555W}~\gsim~22.5$ mag. Lindsay 91, on the other hand,

913: exhibits clear signs of luminosity segregation for most of the magnitude

914: range of interest.

915: 

916: The conversion of an observational LF in a given passband {\it i},

917: $\phi (M_i)$, to its associated MF, $\xi (m)$, is not as

918: straightforward as often assumed (see, e.g., de Grijs et al. 2002b).

919: The differential present-day stellar LF, d{\it N}/d$\phi (M_i)$, i.e.

920: the number of stars in the absolute-magnitude interval $[M_i,M_i+{\rm

921: d}M_i]$, and the differential present-day MF, d{\it N}/d$\xi(m)$, i.e.

922: the mass in the corresponding mass interval $[m,m+dm]$, are

923: related through d$N = -\phi(M_i) dM_i = \xi(m) dm$, and

924: therefore

925: \begin{equation}

926: \label{MLrelation.eq}

927: \phi(M_i) = -\xi(m) {dm \over dM_i} .

928: \end {equation}

929: 

930: Thus, to convert an observational LF into a reliable MF, one needs to

931: have an accurate knowledge of the appropriate ML -- or

932: mass--absolute-magnitude -- relation, d{\it m}/d$M_i$ (for a detailed

933: discussion, see de Grijs et al. 2002b). In fact, it is the {\it slope}

934: of the ML relation at a given absolute magnitude that determines the

935: corresponding mass, which is therefore quite model dependent. This has

936: been addressed in detail by, e.g., D'Antona \& Mazzitelli (1983), Kroupa

937: et al. (1990, 1993), Elson et al. (1995) and Kroupa \& Tout (1997).

938: Given the non-linear shape of the ML relation (de Grijs et al. 2002b)

939: and the small slope at the low-mass end, any attempt to model the ML

940: relation by either a polynomial fit or a power-law dependence will yield

941: intrinsically unreliable MFs (cf. Elson et al. 1995, Chabrier \& M\'era

942: 1997), in particular in the low-mass regime. This model dependence is

943: clearly illustrated by, e.g., Ferraro et al. (1997) and de Grijs et al.

944: (2002b), who compared the MFs for their sample clusters derived from a

945: variety of different ML relations at that time available in the

946: literature. Nevertheless, for the sake of our discussion on mass

947: segregation in our two sample clusters, we only need to consistently

948: apply {\it the same} ML relation to the clusters' annular LFs; any

949: differences in the resulting MF shapes will then be due to intrinsic

950: differences in the stellar mass distributions as a function of radius in

951: the clusters.

952: 

953: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

954: \begin{figure*}[t!]

955: \epsscale{1.15} 

956: \plotone{figure8.eps}

957: \caption{Completeness-corrected LFs of Kron~1 (left) and Lindsay~91

958: (right) within different annuli around the center of the

959: clusters. These LFs clearly show how difficult it is to derive a

960: single slope for the entire LF to compare the results for different

961: clusters. Specific parts of the LF can, however, be approximated by a

962: single power law. The LF of Kron~1 does not show any prominent

963: dependence on the radial distance. Lindsay~91, however,

964: shows a clear radial dependence of its LF in the magnitude range

965: 22.0~\lsim~$m_{\rm F555W}$/mag~\lsim~25.0. One should only consider

966: the LFs in the magnitude range in between the vertical dotted lines,

967: as the brightest stars are affected by small-number statistics and the

968: presence of a red clump (RC), while the lowest-luminosity stars suffer

969: from significant statistical incompleteness. \label{f:lfs}}

970: \end{figure*}

971: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

972: 

973: In addition, the true structure of the CMD above the MS turn-off is very

974: complex. Stellar populations of different masses overlap in

975: color-magnitude space, so that unambiguous mass determination from

976: isochrone fits, for the few dozen stars populating these areas in each

977: cluster, is highly model-dependent. On the other hand, while different

978: evolutionary models differ in terms of absolute calibration of the

979: mass-luminosity relation, they show a quite good agreement in terms of

980: relative mass distribution in well established evolutionary phases. In

981: any case, in a differential analysis such as presented here, the

982: uncertainties involved in the mass determinations of the evolved stars

983: are too large and systematic, so that we cannot include these stars in

984: our MF analysis. Therefore, we used all stars below the MS down to the

985: 50\% completeness limit of the data. After applying the appropriate

986: distance modulus to obtain absolute magnitudes, for our

987: luminosity-to-mass conversions we used the updated Padova isochrones

988: (Marigo et al. 2008; available from {\tt\small

989: http://stev.oapd.inaf.it/cmd}) for the appropriate {\sl HST} photometric

990: system, the metallicity and age of the cluster, and a Kroupa (1998) IMF

991: corrected for the presence of a binary population, as suggested for use

992: with the web interface. For the metallicity and age of both Kron 1 and

993: Lindsay 91 we used $Z = 0.001$ (where Z$_\odot = 0.020$) and an age of

994: 6.3 Gyr; their distance moduli are $m-M = 19.34$ mag and $m-M = 19.10$

995: mag, respectively. These physical parameters are derived from fitting

996: the isochrone model that best matches the observed features of the CMD

997: of each cluster.

998: 

999: The MS turn-off magnitude in both Kron 1 ($m_{\rm F555W} \simeq 22.3$

1000: mag) and Lindsay 91 ($m_{\rm F555W} \simeq 22.1$ mag) corresponds to a

1001: mass of $\log(m_\ast/{\rm M}_\odot) \simeq -0.009$, while the 50\%

1002: completeness limits correspond to $\log(m_\ast/{\rm M}_\odot) = -0.2$

1003: and $-0.17$ for these clusters, respectively. In Fig. \ref{f:mfs} we

1004: show the derived MF slopes as a function of cluster radius for our

1005: adopted ML conversion in the full mass range; the canonical Salpeter

1006: slope would be $\alpha = -2.35$ in this representation. The adopted

1007: radial ranges for our annular MFs are indicated by the horizontal

1008: `error' bars. The vertical error bars represent the formal uncertainty

1009: in the fits. Although the uncertainties are large, both clusters seem to

1010: be affected by mass segregation, with Lindsay 91 being most obviously

1011: mass segregated, out to at least $r \sim 80''$.

1012:  

1013: \subsection{Effective Radii of the Clusters}

1014: 

1015: We calculated the effective radii of the stars in different magnitude

1016: ranges for both clusters. Magnitude bins of various sizes were tested

1017: and we found that a reasonable bin size, which provides good

1018: statistics and allows for the detailed observation of the dependence

1019: of $r_{\rm eff}$ on magnitude, is 0.5~mag. We calculated the effective

1020: radii of the stars as a function of magnitude in the F814W filter. The

1021: relations derived between the computed effective radius and the

1022: corresponding magnitude bin for both Kron~1 and Lindsay~91 are shown

1023: in Fig. \ref{reffvsmag2}. It is shown that, in general, the effective

1024: radius behaves as a function of magnitude for both clusters, thus

1025: providing clear indications of stellar stratification. While this

1026: behavior is more definite for the fainter stars, where it is seen that

1027: for both clusters the effective radii of these stars are larger, the

1028: uncertainties in the $r_{\rm eff}$ calculation for the brightest stars

1029: are rather large due to small-number statistics, and therefore no

1030: definite trend for the bright stars can be derived.

1031: 

1032: From the $r_{\rm eff}$ vs. magnitude plots in Fig.~\ref{reffvsmag2}

1033: one may conclude that the clusters exhibit stellar stratification but

1034: to a different {\sl degree}. Moreover, from the plots in

1035: Fig.~\ref{reffvsmag2} one can derive the degree of stratification of

1036: the clusters in terms of the brightness range of the segregated stars,

1037: as well as the effective radius within which they are

1038: confined. Considering that a steeper slope of the relation $r_{\rm

1039: eff}(m_{\rm F814W})$ represents a higher degree of stratification, we

1040: conclude that Lindsay~91 is more strongly segregated than Kron~1.

1041: 

1042: If we want to parameterize the differences between the clusters and

1043: obtain more quantitative results, we should use the two primary output

1044: parameters of our method, i.e., the {\sl magnitude} and the {\sl

1045: effective radius} of segregation.  Both represent the limit (in

1046: magnitude range and radius, respectively) beyond which the slope of

1047: $r_{\rm eff}(m_{\rm F814W})$ changes significantly.  However, as can be

1048: seen in Fig.~\ref{reffvsmag2}, $r_{\rm eff}$ is not a monotonic function

1049: of brightness. Specifically, for Kron~1, while for stars with $m_{\rm

1050: F814W}$~\lsim~19~mag there is a trend of $r_{\rm eff}$ to become larger

1051: for fainter magnitudes, this trend cannot be confirmed statistically due

1052: to large uncertainties. For fainter magnitudes, down to $\sim$~20.5~mag,

1053: the relation $r_{\rm eff}$ vs. $m_{\rm F814W}$ shows fluctuations and no

1054: specific trend. For even fainter stars, with $m_{\rm

1055: F814W}$~\gsim~21~mag, the slope of this relation shows a definite

1056: steepening as a clear indication that these stars are indeed segregated.

1057: However, this slope is still quite shallow, providing evidence of a {\sl

1058: low degree of segregation}. A comparison between the $r_{\rm eff}$ vs.

1059: $m_{\rm F814W}$ relation of Fig.~\ref{reffvsmag2} for Kron~1 and the LFs

1060: of the cluster constructed for different distances from its center

1061: (Fig.~\ref{f:lfs}), shows that $r_{\rm eff}$ seems to be a function of

1062: brightness for the entire observed magnitude range, but statistics do

1063: not allow the verification of a definite relation. Specifically,

1064: Fig.~\ref{f:lfs} shows that the LF slope seems to be distance-dependent

1065: for the full extent of the cluster and for stars within the entire

1066: observed brightness range.  Indeed, a functional relation between

1067: $r_{\rm eff}$ and brightness seems to exist in Fig.~\ref{reffvsmag2} for

1068: the entire cluster, with $r_{\rm eff}\simeq$~0.7\arcmin\ for $m_{\rm

1069: F814W} \sim$~17~mag, but this trend is not supported statistically. The

1070: small numbers of the brightest stars allow a solid statistical

1071: interpretation of this plot only for stars of $m_{\rm

1072: F814W}$~\gsim~21~mag. For these stars the LF slope becomes steeper

1073: outwards. The advantage of our method lies in the fact that we are able

1074: to define in a direct manner the distance of segregation for these

1075: stars. This distance, derived from Fig.~\ref{reffvsmag2}, is $r_{\rm

1076: eff}\simeq$~0.8\arcmin.

1077: 

1078: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

1079: \begin{figure}[t!]

1080: \epsscale{1.} 

1081: \plotone{figure9.eps}

1082: \caption{Radial dependence of the slopes of the differential MFs of

1083: Kron~1 (top) and Lindsay~91 (bottom). This dependence provides a clear

1084: indication of mass segregation of the stars on the lower MS (below the

1085: turn-off) for both clusters, and particularly for Lindsay 91. However,

1086: the uncertainties in the counting process, the model dependence for

1087: construction of the MF, and the selection of specific radial distances

1088: whithin which the MF is constructed for the analysis of stellar

1089: stratification introduce additional uncertainties, which work their

1090: way through to the results of this analysis.\label{f:mfs}}

1091: \end{figure}

1092: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1093: 

1094: For Lindsay~91 a more obvious trend of $r_{\rm eff}$ as a function of

1095: brightness can be seen for almost the entire observed magnitude range,

1096: $m_{\rm F814W}$~\gsim~18.5~mag.  Also in this cluster, no clear

1097: correlation between $r_{\rm eff}$ and $m_{\rm F814W}$ is observed for

1098: the brighter stars due to their small numbers. The relation between

1099: $r_{\rm eff}$ and $m_{\rm F814W}$ can be clarified easily, however,

1100: for stars with $m_{\rm F814W}$~\gsim~21~mag. It is interesting to note

1101: that from the comparison of the relation of $r_{\rm eff}$ vs. $m_{\rm

1102: F814W}$ of Fig.~\ref{reffvsmag2} for Lindsay~91 with the corresponding

1103: LFs for different distances from the cluster center, shown in

1104: Fig.~\ref{f:lfs}, we see that both plots agree that significant

1105: segregation is observed for stars with $m_{\rm

1106: F814W}$~\gsim~21~mag. However, specific radial distances were selected

1107: for the LFs of Fig.~\ref{f:lfs}, and therefore we cannot define an

1108: accurate {\sl distance of segregation} for these stars, except that

1109: segregation occurs at distances $r$~\gsim~40\arcsec. With our new

1110: method we can accurately define the distance of segregation.  From

1111: Fig.~\ref{reffvsmag2} we derive that this distance corresponds to the

1112: effective radius of $r_{\rm eff} \sim$~0.72\arcmin~$\equiv$~43\arcsec,

1113: which agrees very well with the results from the LFs method, but only

1114: because of the arbitrary selection of this distance for this method.

1115: 

1116: It should be noted that the relation of $r_{\rm eff}$ vs. magnitude

1117: derived for the MS stars (below the turn-off) is the most reliable

1118: indicator of the degree of segregation of the studied clusters. In the

1119: case of this relation for the brighter evolved stars (above the

1120: turn-off) the situation is more complicated because the mass of these

1121: stars changes very slowly with magnitude. Specifically, according to the

1122: evolutionary models adopted here (Marigo et al. 2008) and the isochrone

1123: of 6.3~Gyr, the difference in mass of stars evolving from the base of

1124: the RGB to the AGB is much less than 0.01~M{\solar}, and thus the

1125: corresponding values of $r_{\rm eff}$ refer to extremely narrow ranges

1126: of mass per magnitude bin. As a consequence different magnitude bins

1127: correspond to roughly the same mass and the relation of $r_{\rm eff}$

1128: vs. magnitude for these stars provides indications of stratification of

1129: of stars of different magnitude but nearly identical mass. Naturally,

1130: this effect has consequences on the detection of {\sl mass segregation}.

1131: In the case of the relation of $r_{\rm eff}$ vs. mass, the fact that the

1132: evolved stars have small mass differences can improve the number

1133: statistics, since a larger number of stars will correspond to the same

1134: mass-bins. Indeed, considering that according to the isochrone model all

1135: stars brighter than the turn-off have masses around 1~M{\solar}, if we

1136: group together all these stars, then we should combine the first 10

1137: magnitude bins shown in the plots of Fig.~\ref{reffvsmag2} into one. The

1138: corresponding values of $r_{\rm eff}$ are 0.80\arcmin~$\pm$~0.06\arcmin\

1139: for Kron~1 and 0.88\arcmin~$\pm$~0.07\arcmin\ for Lindsay~91; both

1140: values are quite consistent with the derived trends.

1141: 

1142: In general, based on the above, with the method for detecting and

1143: characterizing the effect of stellar stratification proposed here we can

1144: distinguish between different degrees of segregation among star clusters

1145: and compare our results from one cluster to another. The values of

1146: $r_{\rm eff}$ derived by using only MS stars (below the turn-off) for

1147: the studied clusters should be considered as the best tracers of mass

1148: segregation.

1149: 

1150: %%%%%%%%%%%%%%%%%%%%%%%%%%%% FIGURE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

1151: \begin{figure*}[t!]

1152: \epsscale{1.15} 

1153: \plotone{figure10.eps}

1154: \caption{Estimated effective radii of stars within different magnitude

1155: ranges versus the corresponding mean magnitude for both

1156: clusters.\label{reffvsmag2}}

1157: \end{figure*}

1158: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1159: 

1160: \section{Conclusions}

1161: 

1162: We present a robust new method for the diagnosis of stellar

1163: stratification in star clusters. This method uses the {\sl effective

1164: radius}, which represents the observed Spitzer radius as estimated for

1165: various stellar groups of different magnitude (or mass) ranges in the

1166: same system. This radius is an important dynamical parameter, which

1167: approximately describes the area of gravitational influence of a given

1168: stellar group in the cluster. The proposed method detects stellar

1169: stratification in star clusters: if a cluster is segregated, then the

1170: effective radii for different stellar groups should be different from

1171: one group to another. A relation between the effective radius of every

1172: group and the corresponding selected magnitude (or mass) range can be

1173: established to consistently characterize any stratification. As a result,

1174: comparison among the results obtained for different clusters

1175: can be made.

1176: 

1177: Stellar stratification in star clusters is usually quantified based on

1178: either the surface density profiles of stars in different magnitude

1179: ranges or the cluster LF (or MF) at different radial distances from

1180: its center. Using the first method one may observe the magnitude

1181: (mass) limit for the segregated stars and using the second the radial

1182: distance where they are observed to be segregated. Both of these

1183: results can be derived using the single application of the method

1184: proposed here. Furthermore, to apply the `classical' methods for the

1185: detection of mass segregation, certain complicated fitting procedures

1186: must be applied, leading to model-dependent results. On the other

1187: hand, our method is much more straightforward in its application, and

1188: more precise in its results. This is mostly so because it is {\sl not}

1189: necessary to (i) introduce any significant assumptions to estimate the

1190: cluster parameters, and (ii) apply any fit to parameter correlations

1191: and subsequently test the relation between the model output and the

1192: observables. Both of these steps must be applied in the classical

1193: approaches to characterize stellar stratification.

1194: 

1195: We present the application of this new diagnostic tool to two

1196: mass-segregated clusters in the SMC, observed with {\sl HST}/ACS. The

1197: application of the new method to space-based observations helped us to

1198: establish a scheme for the comparison between the results for different

1199: clusters and to define the important parameters required to characterize

1200: stratification. These parameters are the {\sl magnitude} and {\sl radius

1201: of segregation}, which specify the limits beyond which segregation

1202: becomes stronger and more apparent. In addition, one may quantify the

1203: differences between clusters through the {\sl slope of the relation

1204: between the effective radius and the corresponding magnitude range},

1205: which can be used as an indicator of the degree of segregation of each

1206: cluster.

1207: 

1208: In each of the clusters observed with the {\sl HST}, considered here, it

1209: was found that segregation is apparent for stars over the full observed

1210: brightness range, but small-number statistics do not allow the detection

1211: of a definite trend between $r_{\rm eff}$ and magnitude for several of

1212: the brightest stellar groups. A comparable {\sl radius of segregation}

1213: is observed for stars of $m_{\rm F814W}$~\gsim~21~mag in both clusters,

1214: but the {\sl degree of segregation} seems to differ from one cluster to

1215: the other, with Lindsay~91 exhibiting a steeper function of $r_{\rm

1216: eff}(m_{\rm F814W})$ than Kron~1.

1217: 

1218: The continuous increase of $r_{\rm eff}$ as a function of brightness

1219: away from the radius of segregation up to the observed limits of the

1220: clusters shows that the fainter stars are distributed throughout a

1221: larger volume, an observation reminiscent of the `evaporation' of star

1222: clusters. Still, these results are based on only two clusters and can

1223: therefore only be used rather qualitatively.

1224: 

1225: We conclude that the method we present here is simpler and more

1226: accurate than previously developed methods. This is based, first and

1227: foremost, on the more straightforward way in which the effective

1228: radius is estimated and, secondly, on avoiding any model-dependent

1229: fits. Instead, a direct correlation between the cluster's observables

1230: is sufficient to exhibit any stellar stratification and to

1231: parameterize its significance.

1232: 

1233: \acknowledgements

1234: 

1235: D. A. Gouliermis kindly acknowledges the support of the German Research

1236: Foundation (Deu\-tsche For\-schungs\-ge\-mein\-schaft - DFG) through the

1237: research grant GO~1659/1-1. We thank K. S. de~Boer for his contribution

1238: to an earlier version of this paper, and R. Korakitis for his comments

1239: regarding the projection of a spherical system. Based on observations

1240: made with the NASA/ESA {\em Hubble Space Telescope}, obtained from the

1241: data archive at the Space Telescope Science Institute. STScI is operated

1242: by the Association of Universities for Research in Astronomy, Inc. under

1243: NASA contract NAS 5-26555.

1244: 

1245: %Facilities: HST

1246: 

1247: %\newpage

1248: 

1249: \begin{references}

1250: 

1251: \reference{} Bonnell, I. A., \& Davies, M. B. 1998, MNRAS 295, 691

1252: 

1253: \reference{} Bonatto, C., \& Bica, E.\ 2007, MNRAS, 377, 1301

1254: 

1255: \reference{} Brandl, B., Sams, B. J., Bertoldi, F., et al. 1996, ApJ, 466,

1256: 254

1257: 

1258: \reference{} Brocato, E., Di Carlo, E., \& Menna, G.\ 2001, A\&A, 374, 523

1259: 

1260: \reference{} Brocato, E., Castellani, V., Di Carlo, E., Raimondo, G., \&

1261: Walker, A.~R.\ 2003, AJ, 125, 3111

1262: 

1263: \reference{} Chabrier, G., \& M\'era, D.\ 1997, A\&A, 328, 83

1264: 

1265: \reference{} D'Antona, F., \& Mazzitelli, I.\ 1983, A\&A, 127, 149

1266: 

1267: \reference{} de Grijs, R., Johnson, R.~A., Gilmore, G.~F., \& Frayn, C.~M.\

1268: 2002a, MNRAS, 331, 228

1269: 

1270: \reference{} de Grijs, R., Gilmore, G.~F., Johnson, R.~A., \& Mackey, A.~D.\

1271: 2002b, MNRAS, 331, 245

1272: 

1273: \reference{} de Grijs, R., Gilmore, G.~F., Mackey, A.~D., Wilkinson, M.~I.,

1274: Beaulieu, S.~F., Johnson, R.~A., \& Santiago, B.~X.\ 2002c, MNRAS, 337,

1275: 597

1276: 

1277: \reference{} Elson, R. A. W., Fall, S. M., \& Freeman, K. C. 1987, ApJ, 323,

1278: 54

1279: 

1280: \reference{} Elson, R. A. W., Gilmore, G. F., Santiago, B. X., \& Casertano,

1281: S.\ 1995, AJ, 110, 682

1282: 

1283: \reference{} Fischer, P., Pryor, C., Murray, S., Mateo, M., \& Richtler, T.\

1284: 1998, AJ, 115, 592

1285: 

1286: \reference{} Flower, P.~J., Geisler, D., Olszewski, E.~W., \& Hodge, P.\

1287: 1980, ApJ, 235, 769

1288: 

1289: \reference{} Dolphin, A.~E., et al. 2001, ApJ, 562, 303 

1290: 

1291: \reference{} Elson, R. A. W., Gilmore, G. F., \& Santiago, B. X. 1997, MNRAS,

1292: 289, 157

1293: 

1294: \reference{} Fischer, P., Pryor, C., Murray, S., Mateo, M., \& Richtler, T.\

1295: 1998, AJ, 115, 592

1296: 

1297: \reference{} Gouliermis, D., Kontizas, M., Korakitis, R., et al. 2000, AJ,

1298: 119, 1737

1299: 

1300: \reference{} Gouliermis, D., Keller, S.~C., Kontizas, M., Kontizas, E., \&

1301: Bellas-Velidis, I.\ 2004, A\&A, 416, 137

1302: 

1303: \reference{} Hillenbrand, L. A., \& Hartmann, L. W. 1998, ApJ, 492, 540

1304: 

1305: \reference{} Holtzman, J. A., et al. 1999, AJ, 118, 2262

1306: 

1307: \reference{} Johnson, J.~A., Bolte, M., Stetson, P.~B., Hesser, J.~E., \&

1308: Somerville, R.~S.\ 1999, ApJ, 527, 199

1309: 

1310: \reference{} Keller, S.~C., Bessell, M.~S., \& Da Costa, G.~S.\ 2000, AJ,

1311: 119, 1748

1312: 

1313: \reference{} King, I. R. 1965, ApJ, 142, 387

1314: 

1315: \reference{} King, I. R. 1966, AJ, 71, 64

1316: 

1317: \reference{} King, I.~R., Sosin, C., \& Cool, A.~M.\ 1995, ApJL, 452, L33

1318: 

1319: \reference{} Keller, S.~C., Bessell, M.~S., \& Da Costa, G.~S.\ 2000, AJ,

1320: 119, 1748

1321: 

1322: \reference{} Kontizas, M., Hatzidimitriou, D., Bellas-Velidis, I.,

1323: Gouliermis, D., Kontizas, E., \& Cannon, R.~D.\ 1998, A\&A, 336, 503

1324: 

1325: \reference{} Kron, G. E.\ 1956, PASP, 68, 125 

1326: 

1327: \reference{} Kroupa, P.\ 1998, MNRAS, 298, 231

1328: 

1329: \reference{} Kroupa, P., \& Tout, C. A.\ 1997, MNRAS, 287, 402

1330: 

1331: \reference{} Kroupa, P., Tout, C. A., \& Gilmore, G. F.\ 1990, MNRAS, 244,

1332: 76

1333: 

1334: \reference{} Kroupa, P., Tout, C. A., \& Gilmore, G. F.\ 1993, MNRAS, 262,

1335: 545

1336: 

1337: \reference{} Lightman, A. P. \& Shapiro, S.L., 1978, Rev. Mod. Phys. 50, 437

1338: 

1339: \reference{} Lindsay, E. M.\ 1958, MNRAS, 118, 172 

1340: 

1341: \reference{} Mackey, A.~D., \& Gilmore, G.~F.\ 2003a, MNRAS, 338, 120 

1342: 

1343: \reference{} Mackey, A.~D., \& Gilmore, G.~F.\ 2003b, MNRAS, 340, 175 

1344: 

1345: \reference{} Mackey, A.~D., \& Gilmore, G.~F.\ 2004, MNRAS, 352, 153

1346: 

1347: \reference{} Mackey, A.~D., Payne, M.~J., \& Gilmore, G.~F.\ 2006, MNRAS,

1348: 369, 921

1349: 

1350: \reference{} Ma{\'{\i}}z Apell{\'a}niz, J., \& {\'U}beda, L.\ 2005, ApJ,

1351: 629, 873

1352: 

1353: \reference{} Marigo, P., Girardi, L., Bressan, A., Groenewegen, M.~A.~T.,

1354: Silva, L., \& Granato, G.~L.\ 2008, A\&A, 482, 883

1355: 

1356: \reference{} Meylan, G. \& Heggie, D.C. 1997, A\&A Rev. 8, 1

1357: 

1358: \reference{} Murray, S. D., \& Lin, D. N. C. 1996, ApJ 467, 728

1359: 

1360: \reference{} Pandey, A.~K., Mahra, H.~S., \& Sagar, R.\ 1992, Bulletin of the

1361: Astronomical Society of India, 20, 287

1362: 

1363: \reference{} Plummer, H. C. 1911, MNRAS, 71, 460

1364: 

1365: \reference{} Rochau, B., Gouliermis, D.~A., Brandner, W., Dolphin, A.~E., \&

1366: Henning, T.\ 2007, ApJ, 664, 322

1367: 

1368: \reference{} Sagar, R., Myakutin, V.~I., Piskunov, A.~E., \& Dluzhnevskaya,

1369: O.~B.\ 1988, Bulletin of the Astronomical Society of India, 16, 87

1370: 

1371: \reference{} Scaria, K.~K., \& Bappu, M.~K.~V.\ 1981, Journal of Astrophysics

1372: and Astronomy, 2, 215 (Corrigendum: 1981, Journal of Astrophysics and

1373: Astronomy, 2, 439)

1374: 

1375: \reference{} Santiago, B., Beaulieu, S., Johnson, R., \& Gilmore, G.~F.\

1376: 2001, A\&A, 369, 74

1377: 

1378: \reference{} Sirianni, M., Nota, A., De Marchi, G., Leitherer, C., \&

1379: Clampin, M.\ 2002, ApJ, 579, 275

1380: 

1381: \reference{} Smecker-Hane, T. A., Cole, A. A., Gallagher, J. S., \& Stetson,

1382: P. B. 2002, ApJ, 566, 239

1383: 

1384: \reference{} Spitzer, L. Jr. 1958, ApJ, 127, 17 

1385: 

1386: \reference{} Spitzer, L. Jr. 1969, ApJ, 158, L139 

1387: 

1388: \reference{} Stanghellini, L., Villaver, E., Shaw, R.~A., \& Mutchler, M.\

1389: 2003, ApJ, 598, 1000

1390: 

1391: \reference{} Subramaniam, A., Sagar, R., \& Bhatt, H. C. 1993, A\&A, 273, 100

1392: 

1393: \reference{} Testa, V., Ferraro, F.~R., Chieffi, A., Straniero, O., Limongi,

1394: M., \& Fusi Pecci, F.\ 1999, AJ, 118, 2839

1395: 

1396: \reference{} Xin, Y., Deng, L., de~Grijs, R., Mackey, A.~D., \& Han Z. 2008,

1397: MNRAS, 384, 410

1398: 

1399: %\reference{} Shapley H., Lindsay E. M. 1963, IrAJ., 6, 74

1400: 

1401: \end{references}

1402: 

1403: \end{document}

1404: 

1405: 

1406: